[
https://issues.apache.org/jira/browse/OPENNLP-261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13091097#comment-13091097
]
Tommaso Teofili commented on OPENNLP-261:
-----------------------------------------
I managed to create also some integration tests for the collection reader and
the cas consumer I wrote; just one needs to remove the XmiCasDB and index
directories that get created after the test.
The only thing I had to change was the LuceneSearchService which has hardcoded
boolean parameter (set to false) for the creation of the index while I had to
set it true for the sake of test. Obviously if that could be made configurable
(in the LSS initialization I think) I think it would be good.
Another thing I'd like to review is how to automatically create the casId
inside the CAS Consumer from the CAS itself, do you already have ideas on how
to do that?
Currently I do :
{code}
String.valueOf(cas.getDocumentAnnotation().getCoveredText().hashCode())
{code}
as the DocumentAnnotation should be always present in the CAS and Strings have
proper hashcode() implementation (not deriving from Object).
However this is just one way ;)
> Wirte Analysis Engines to read and write CAS from/to the Corpus Server
> -----------------------------------------------------------------------
>
> Key: OPENNLP-261
> URL: https://issues.apache.org/jira/browse/OPENNLP-261
> Project: OpenNLP
> Issue Type: Task
> Components: Corpus Server
> Reporter: Joern Kottmann
>
> We need to write an Analysis Engine to read CASes from the Corpus Server, and
> one to write analysis results back into the Corpus Server.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
